This relates generally to graphics processing.
In graphics processing, the depths of objects within a depiction may be coded. Advantageously, the coding used for the depth information is compressed so that the depth buffer size is reduced, saving costs.
In some cases, a single depth value is used for each pixel. In other cases, called multi-sampled anti-aliasing, multiple depth values are used for each pixel. In this case, there may be some number of samples per pixel and there may be a depth value associated with each of those samples.
The problem that arises with depth compression, for multi-sampled anti-aliasing, is that the size of the depth buffer increases with increasing numbers of samples. Since more samples is generally desirable, and because the cost of a depth buffer increases with increasing size, there is a need for ways to store information with increasing numbers of samples without excessively increasing the depth buffer cost.